Fullscreen HTML5Med Res (??MB)HTML4

RADIO CONTROL BY BOB YOUNG Transmitter interference on the 36MHz band In previous months we have discussed the possibility of transmitter interference on the 36MHz band. This month, we pres­ent a series of measurements which finally demonstrates an area where FM is actually superior to AM. What’s this? Is Bob Young about to recant and admit that FM has been superior to AM all along? Well, not quite. But I have been able to demonstrate and measure practical cases of interfer­ence between transmitters on the 36MHz band for both AM and FM transmitters and the results are very interesting. In the February 1997 column we warned of the possibility of transmitter intermodulation causing interference when two trans­mitters separated by 455kHz were operated simultaneously on the 36MHz band. Then in March 1997 we presented solutions aimed at preventing this problem. This month we look at practical situations wherein this form of interference may arise if the correct operational pro­cedures are not adhered to. How serious is the problem? For those who have missed previous Fig.1: the spectrum plot of the mixer output, before the filter­ing, of an AM receiver operating on 36MHz from a transmitter on the correct frequency (channel 631, 36.310MHz). The fundamental output is at 455kHz. Note that there is some jitter in the spec­trum plot due to the frequency shift keying of the transmitter. 72  Silicon Chip articles, the problem we are discussing is the transmitter inter­ mod­ ulation component that will arise in the mixer of any single conversion receiver regard­ less of frequency, when two transmitters separated by 455kHz are operated simultaneously. As the 36MHz band is the only Australian R/C band wide enough to accommodate transmitters 455kHz apart, this problem is exclusive to that band. Having discovered this potentially serious problem, it was up to me to make more measurements to define whether it was going to be a real problem on the operating field. With that in mind I gathered a representative batch of modern R/C equipment of various brands with the help of several trade houses, together with receivers of various brands from Fig.2: the intermodulation product of two FM transmitters sepa­rated by 460kHz operating at the same distance from the receiver. In this instance the correct transmitter has been turned off for the sake of clarity. Note that the amplitude of the two signals is actually slightly higher than the original shown in Fig.1. Fig.3: this scope plot shows the normal output of a typical Japanese FM receiver at a test point after the detector, squelch and noise filtering, with the primary and one of the intermodu­lating pair of transmitters operating simultan­ eously. The receiv­er is on 36.370MHz and the other trans­ mitter is on 36.070MHz. Note that there is no sign of any interference. my own stock. Then it was into serious measurements in order to get a better grasp of the situation. No mixer output Modern FM receivers present us with a problem here as the output of the mixer is not easily accessible. This is because almost all models use an IC receiver chip. Therefore we had to cheat in this respect. Fig.1 shows a spectrum plot of the mixer output, before the filtering, of an AM receiver operating on 36MHz from an FM transmitter on the correct frequency (Channel 631, 36.310MHz). While this method might seem invalid, the method of modula­tion does not matter at this point, as we are only looking at the raw, undifferentiated 455kHz mixer component. Note the amplitude of the 455kHz component. By the way, there is only one spike at 455kHz; the double spike in the photo is due to jitter in the spectrum plot due to the frequency shift keying of the transmitter. Fig.2 shows the intermodulation product of two FM transmit­ters separated by 460kHz operating at the same distance from the receiver. In this instance the correct transmitter has been turned off for the sake of clarity. Note that the amplitude of the two signals is actually slightly higher than Fig.4: this scope plot shows the output of the same receiver (as Fig.3) at the same test point but with the primary (wanted) transmitter switched off and an unmodulated signal generator on x36.075MHz and a transmitter on 36.530MHz. Here we are generating an exact 455kHz intermodulation (difference) product from two interfering transmitters. the original shown in Fig.1. Now the really important point to note is that this receiv­er is tuned to 36.310MHz, which is nowhere near the frequencies of the two offending transmitters. So here we have proof of the central point of this series of articles: two transmitters operating simultaneously and sepa­rated by 450kHz or 460kHz will generate a strong 450kHz or 460kHz component in the mixers of every single conversion receiver operating on the 36MHz band. This is regardless of the frequency of the receivers and the frequencies of the intermodulating pair of transmitters! Yes, you understood perfectly. All 59 receivers will be affected simultaneously by just one pair of inter­ This photo shows some of the equipment used to make the measurements discussed in this month’s article. Not shown are the spectrum analyser and some of the receivers. May 1997  73 Fig.5: this shows the same setup as before but with the signal generator at 36.071MHz, just 4kHz away from the 455kHz ideal. Note how distorted the signal has become, indicating severe attenuation in the receiver bandwidth filter. modulating transmitters. As a matter of interest, I checked to see if a pair of 36MHz transmitters would interfere with 29MHz receivers and fortunately they did not. So this potential transmitter interference problem is not just a theory. It does exist and is easily measurable. Two trans­ mitters separated by 450- 460kHz will generate a powerful inter­ modulation component in the mixers of single conversion receiv­ers. The level of this component can equal or exceed the primary transmitter signal, depending upon a whole range of factors. The most obvious factor is the relative signal strength ratios between the primary transmitter and the inter­modulating pair. This is a most important factor in R/C operations and we will examine this later. More subtle factors include mixer compres­sion and bandwidth of the mixer output. Mixer compression arises due to the fact that the mixer can only handle a finite signal level. As more signals arrive at the mixer the amplitude of each component is reduced accordingly. Theoretically, if the intermodulation product is 455kHz, the mixer bandwidth should not play any part in this discussion. However, in the real world the intermodulation product is not 455kHz but 450kHz or 460kHz, 74  Silicon Chip Fig.6: this shows the same receiver with the same two transmit­ters operating but with the third transmitter also switched on. This transmitter is on 36.530MHz, so the intermodulation compon­ent is 460kHz. Note how disturbed the output has become. While capture has not been achieved, the wanted transmitter has lost control. because of the 10kHz spacing between adjacent channels (see March 1997 issue). So the mixer bandwidth becomes an important factor. 5kHz protection Fig.3 shows the normal output of a typical Japanese FM receiver at a test point after the detector, squelch and noise filtering, with the primary and one of the intermodulating pair of transmitters operating simultaneously. The receiver is on 36.370MHz and the other transmitter is on 36.070MHz. Note that there is no sign of any interference. Fig.4 shows the output of the same receiver at the same test point but with the primary (wanted) transmitter switched off and an unmodulated signal generator on 36.075MHz and a transmit­ter on 36.530MHz. Here we are generating an exact 455kHz inter­ modulation (difference) product. Note that we are getting the perfectly normal output wave­form even though we are generating the control signal from the intermodulation product on a receiver nowhere near the two RF signal sources (36.370MHz). The unmodulated signal generator is necessary to generate a normal waveform. If two modulated trans­mitters were used the resultant composite modulation would drive the servos wild. Now we arrive at the interesting bit. Fig.5 shows the same setup as before but with the signal generator at 36.071MHz, just 4kHz away from the 455kHz ideal. Note how distorted the signal has become, indicating severe attenuation in the receiver band­width filter. This is the saving grace in this whole affair. Three distinct and separate factors have come together in the real world to make practical operation a reasonably safe proposition. First, due the fact that the channels are spaced every 10kHz and that the IF is 455kHz, the intermodula­tion product falls midway between two channels; ie, 5kHz away from the channels on either side. Second, modern receivers have a typical bandwidth of around +5kHz and -7kHz (<at> 40dB) and the attenuation of any signal 5kHz away from 455kHz is such that the genuine 455kHz signal will become dominant. This then leads to the importance of the third factor, “capture effect”, which ensures that only the dominant signal has control. So does transmitter intermodulation present a serious prob­lem in the real world on the 36MHz band? The answer is a reserved no. Why are there reservations? Answer: because of the variations in receiver performance. Can you guarantee that your receiver’s bandwidth is as good as typical Fig.7: capture can occur if the conditions are correct. Here the signal generator is set at 36.075MHz to simulate a transmitter off-frequency or a receiver with a wider than usual bandwidth and the third transmitter is on 36.530MHz. To achieve capture, the primary transmitter has been moved away, thus simulating condi­tions which can be encountered on flying fields. modern receivers (+7kHz, -5kHz)? Can you guarantee that the relative transmitter signal ratios will always favour the wanted transmitter? Let’s look at some of these factors in more detail. Fig.6 shows the same receiver with the same two transmit­ters operating but with the third transmitter also switched on. This transmitter is on 36.530MHz, so the intermodulation compon­ent is 460kHz. Note how disturbed the output has become. While capture has not been achieved, the wanted transmitter has lost control. To achieve this result, the inter­ modulating pair of trans­mitters had to be much closer to the receiver than the primary transmitter. With all three transmitters at equal distances from the receiver, there was no sign of any interference. Capture In the testing done so far on a small batch of imported receivers, results varied from excellent to good. Even different models from the same manufacturer gave different results in regards to capture, as would be expected from normal production tolerances. In most instances, capture was difficult to obtain, requir­ing unrealistic signal ratios – signal ratios that could never be achieved on any R/C field. Fig.8: this shows the result of an AM receiver subjected to an identical level of intermodulation interference as the FM receiv­er in Fig.4. Whilst AM receivers have capture ratios of 100:1 or more, long before capture the signal becomes very disturbed as shown here. In one instance, capture could not be achieved but that receiver just simply stopped working. Again, this was at unrealistic signal levels. Remember here that the receiver has already captured its primary transmitter and in order to take control away from that primary, the interference must exceed the level of the primary signal. The ratio between the interfering signal and the primary signal is known as the capture ratio and is usually in the order of 1-3dB. In simple ratio terms, these correspond to trans­mitter signal ratios of 1.12:1 up to 1.41:1. Now we can see why the 5kHz difference between the inter­modulation product and the primary product is so important. If the signal level of the intermodulation product can be reduced to just below the primary, capture is virtually impossible. Fig.7 shows that capture can occur if the conditions are correct. Here the signal generator is set at 36.075MHz to simu­late a transmitter off-frequency or a receiver with a wider than usual bandwidth. The third transmitter is on 36.530MHz. To achieve capture, the primary transmitter has been moved away, thus simulating conditions which can be encountered on flying fields. Note the ripple on the baseline of the scope trace, indi­cating a strong transmitter still present on the correct frequen­cy. And finally what of the situation that started all of this – two models operating on frequencies 450kHz or 460kHz apart? A quick test indicated that with just two transmitters operating (607, 653), the servos started to jump as the second transmitter (607) came close to the receiver. (653). The same test repeated with a receiver on 637 showed no sign of interference, even with the transmitter antenna touching the receiver antenna. Thus there is still a case for not operating two overlapping frequencies simultaneously, regardless of the foregoing arguments. What does it mean in the field? What it means is that under normal conditions, using FM receivers, there is little likelihood of any interference being experienced as long as safe operating practices are followed. Here I should refer to the discussions and illustrations of the flying field layout published in the July 1995 issue of SILICON CHIP. Fig.9 is reproduced from that article. This depicts the real danger of transmitter inter­modulation in a practical sense. If the intermodulating pair of transmitters are located at the end of the May 1997  75 Fig.9: this diagram is reproduced from the July 1995 issue of SILICON CHIP. It depicts the real danger of transmitter intermod­ulation in a practical sense. Interference is more likely when the controlling transmitter is further away from the receiver. flightline closest to the model and the primary transmitter is situated at the far end of the flightline, then we have the conditions for interference, if not capture. Even mild interference on final approach is enough to result in a damaged model. A similar set of conditions can prevail on glider fields where a pilot may leave the flightline to go down the field to the bungie site during launch. After launch and before the pilot can return to the flightline, the model may pass close to the group of transmitters and thus the intermodulating pair, thereby setting up conditions for interference. Time and time again we return to the absolute necessity for adherence to the correct operational procedures on all R/C fields. Ignore this warning at your peril! What about AM? This leaves us with the final point to discuss in this issue. There have 76  Silicon Chip been rumblings for some time about AM receivers being plagued with interference on 36MHz. The MAAA is considering banning AM on 36MHz as a result. The ubiquitous grapevine attrib­ utes this interference to harmonics from the broadcast FM trans­ missions. I wonder if this problem is due to transmitter inter­ modulation? Fig.8 shows the results of an AM receiver subjected to an identical level of intermodulation interference as the FM receiver in Fig.4. Whilst AM receivers have capture ratios of 100:1 or more, long before capture the signal becomes very dis­turbed as in Fig.8. Thus without capture effect to protect them, AM receivers could suffer badly on 36MHz as long as overlapping pairs of transmitters are allowed to operate. There is no doubt that capture effect, whilst a two-edged sword, does give the FM re­ceiver the edge over AM in this situation. On 29MHz, this situation does not apply and my original remarks regarding AM versus FM still apply. And if overlapping transmissions are stopped, AM should be perfectly safe on 36MHz. Actually this entire series of articles was sparked off some months ago as a result of “experts” in a club telling a beginner who was constantly crashing to get rid of his “inferior” 36MHz AM equipment or he would not be allowed to fly in that club. It would be the ultimate irony if it turned out that it was the “superior” FM transmitters causing this poor fellow’s miser­ies! In conclusion, as a result of the uncertainties surrounding the problem of transmitter intermodulation I would recommend that transmitters 450kHz or 460kHz apart not be operated simultaneously on model flying fields. The Silvertone Keyboard provides a simple method of controlling this situation. Acknowledgement I would like to extend my appreciation to Hobby Headquar­ters (NSW) and L. O’Reilly Pty Ltd (SA) for the loan of the equipment used in this SC article. Bob Young is the principal of Silvertone Electronics. Phone/fax (02) 9533 3517.